data augmentation does for generalization: artificially expands training set

How can you teach a computer to see better?

Imagine you're teaching your friend to recognize cats. You show them pictures of cats with different backgrounds, lighting, and angles. But after a while, they start to forget the cats' unique features because everything looks too similar.

To prevent your friend from forgetting, you could show them more varied pictures of cats. This way, they learn to spot cats even if they're in different situations. This is what data augmentation does for machine learning models.

Example

You show your friend 10 pictures of cats, then 10 more with slight changes like a different background or angle. After seeing 20 pictures, they can spot cats in new pictures better.

Remember this

Data augmentation helps machine learning models generalize better by exposing them to more varied examples.

Text adapted from Wikipedia, licensed under CC BY-SA 4.0.

Related concepts

batch size affects generalization: larger batches find sharper minima

Larger batch sizes lead to sharper minima, enhancing generalization by providing more accurate gradient estimates

dropout works as regularization: it approximates an ensemble of subnetworks

Why does turning off neurons randomly help a brainy computer learn better?

Machine learning

Can we teach computers to learn like humans?

gradient accumulation simulates larger batch sizes without more memory

Can you train a machine like you do with a computer?

Adam vs SGD: Adam adapts per-parameter rates, SGD often generalizes better with tuning

Adam adjusts learning rates per-parameter, SGD generalizes better with tuning

General-purpose computing on graphics processing units

Did you know your computer can do more than just compute numbers?

Swipe through 100 ML concepts daily

Open Pocket Polymath